How to install via conda manually?

[jolespin]

Here is your install.sh script:

# install python packages
#python -mpip install -U pip numpy==1.16.4 scipy==1.3.0 networkx==2.3 cffi==1.12.3 biopython==1.73
conda install -y networkx=2.3 cffi=1.12.3 biopython=1.73 numpy=1.16.4 scipy=1.3.0

rm -rf ./pypy
# 1. download portable pypy
wget -c https://bitbucket.org/squeaky/portable-pypy/downloads/pypy-7.1.0-linux_x86_64-portable.tar.bz2

# 2. unzip the compressed file
tar xvf pypy-7.1.0-linux_x86_64-portable.tar.bz2

# 3. remove the compressed file
rm pypy-7.1.0-linux_x86_64-portable.tar.bz2

# 4. install pypy

mv ./pypy-7.1.0-linux_x86_64-portable/ ./pypy
cd pypy
mkdir install_dir
./bin/virtualenv-pypy -p ./bin/pypy ./install_dir

./install_dir/bin/pypy -mpip install -U pip rpython

#pypy/bin/pypy ./bin/find_hit.py &> /dev/null
#./install_dir/bin/pypy ../bin/find_hit.py &> /dev/null
./install_dir/bin/pypy ../bin/find_hit.py

I have a conda environment I've created for this program:

(swiftortho_env) -bash-4.1$ echo $CONDA_PREFIX
/usr/local/devel/ANNOTATION/jespinoz/anaconda3/envs/swiftortho_env

I want to put everything in here but I'm confused on your install script. I get what conda install -y networkx=2.3 cffi=1.12.3 biopython=1.73 numpy=1.16.4 scipy=1.3.0 is doing. I'm a little confused on why portable-pypy is being used instead pip.

I have 2 directories:
/usr/local/devel/ANNOTATION/jespinoz/anaconda3/envs/swiftortho_env/bin
/usr/local/devel/ANNOTATION/jespinoz/anaconda3/envs/swiftortho_env/lib

Can I just move your bin and lib files into these directories?

It would be much easier to have a conda install -c bioconda swiftortho in the future if possible.

[Rinoahu]

1. Why use portable-pypy:
   a. portable-pypy is used to install RPtyhon package. CPython is a very slow language, we used RPython to convert python code to c/c++, which significantly speeds up.
   b. RRpython can be only installed on CPython 2.7 (or below) and PyPY. It is not compatible with Python 3.6 (or above).
   c. SwiftOrtho has removed compatibility with CPython 2.7 because CPython 2.7's official support has been stopped,

2. You can try to move these files to those directories, but I do not test it yet and don't how if it works.

3. Using conda install may be easier but there is an environment issued of RPython that needs to be fixed. If you wanna make it easy to use, you can temporally use Linux alias.

[xiekunwhy]

Also hoping that SwiftOrtho can be installed from conda.

[jolespin]

Also hoping that SwiftOrtho can be installed from conda.

Check out the clustering module of VEBA. It uses MMSEQS2. Not sure what your purpose is but MMSEQS2 was able to achieve what I was trying to do.

https://github.com/jolespin/veba